c語言之傅立葉變化
阿新 • • 發佈:2019-01-01
#include <stdio.h> void StartTimer0(void); void Interrupt_Init(void); void Interrupt_Start(void); #if 0 void main (void) { printf("Start Timer\n"); StartTimer0();//設定定時器暫存器 // Interrupt_Init(); // Interrupt_Start(); printf("while(1)\n"); // while(1); } #endif #include <math.h> #include <stdio.h> #include <stdlib.h> #include <string.h> //精度0.0001弧度 typedef struct//複數型別 { float real; //實部 float imag; //虛部 }complex; #define PI 3.1415926 /////////////////////////////////////////// void conjugate_complex(int n,complex in[],complex out[]); void c_plus(complex a,complex b,complex *c);//複數加 void c_mul(complex a,complex b,complex *c) ;//複數乘 void c_sub(complex a,complex b,complex *c); //複數減法 void c_div(complex a,complex b,complex *c); //複數除法 void fft(int N,complex f[]);//傅立葉變換 輸出也存在陣列f中 void c_abs(complex f[],float out[],int n);//複數陣列取模 void conjugate_complex(int n,complex in[],complex out[]) { int i = 0; for(i=0;i<n;i++) { out[i].imag = -in[i].imag; out[i].real = in[i].real; } } void c_abs(complex f[],float out[],int n) { int i = 0; float t; for(i=0;i<n;i++) { t = f[i].real * f[i].real + f[i].imag * f[i].imag; out[i] = sqrt(t); } } void c_plus(complex a,complex b,complex *c) { c->real = a.real + b.real; c->imag = a.imag + b.imag; } void c_sub(complex a,complex b,complex *c) { c->real = a.real - b.real; c->imag = a.imag - b.imag; } void c_mul(complex a,complex b,complex *c) { c->real = a.real * b.real - a.imag * b.imag; c->imag = a.real * b.imag + a.imag * b.real; } void c_div(complex a,complex b,complex *c) { c->real = (a.real * b.real + a.imag * b.imag)/(b.real * b.real +b.imag * b.imag); c->imag = (a.imag * b.real - a.real * b.imag)/(b.real * b.real +b.imag * b.imag); } #define SWAP(a,b) tempr=(a);(a)=(b);(b)=tempr void Wn_i(int n,int i,complex *Wn,char flag) { Wn->real = cos(2*PI*i/n); if(flag == 1) Wn->imag = -sin(2*PI*i/n); else if(flag == 0) Wn->imag = -sin(2*PI*i/n); } //傅立葉變化 void fft(int N,complex f[]) { complex t,wn;//中間變數 int i,j,k,m,n,l,r,M; int la,lb,lc; /*----計算分解的級數M=log2(N)----*/ for(i=N,M=1;(i=i/2)!=1;M++); /*----按照倒位序重新排列原訊號----*/ for(i=1,j=N/2;i<=N-2;i++) { if(i<j) { t=f[j]; f[j]=f[i]; f[i]=t; } k=N/2; while(k<=j) { j=j-k; k=k/2; } j=j+k; } /*----FFT演算法----*/ for(m=1;m<=M;m++) { la=pow(2,m); //la=2^m代表第m級每個分組所含節點數 lb=la/2; //lb代表第m級每個分組所含碟形單元數 //同時它也表示每個碟形單元上下節點之間的距離 /*----碟形運算----*/ for(l=1;l<=lb;l++) { r=(l-1)*pow(2,M-m); for(n=l-1;n<N-1;n=n+la) //遍歷每個分組,分組總數為N/la { lc=n+lb; //n,lc分別代表一個碟形單元的上、下節點編號 Wn_i(N,r,&wn,1);//wn=Wnr c_mul(f[lc],wn,&t);//t = f[lc] * wn複數運算 c_sub(f[n],t,&(f[lc]));//f[lc] = f[n] - f[lc] * Wnr c_plus(f[n],t,&(f[n]));//f[n] = f[n] + f[lc] * Wnr } } } } #define NUM 512 int main() { int N,i; unsigned short amp[NUM]; complex input[NUM]; int fs = 44100; N = 512; for(i=0;i<N;i++) { input[i].real = 7000*cos(1000*2*(double)PI*((double)i/44100) + (double)3/5*PI) + 200*cos(600*2*(double)PI*((double)i/44100) + (double)1/2*PI); input[i].imag = 0; } fft(N,input); printf("得到的頻譜值為:\n"); for(i=0;i<N/2;i++) { double originamp = sqrt(input[i].real*input[i].real + input[i].imag*input[i].imag); if(i == 0) { amp[i] = originamp/N; } else { amp[i] = originamp/N*2; } if(amp[i] > 20) { printf("f:%.2lf amp:%d\n",(double)i*fs/N,amp[i]); } } return 0; }